Application of DEA-ANN for Best Performance Modeling

The purpose of this study is to present a complementary modeling approach using data envelopment analysis (DEA) and artificial neural network (ANN) as an adaptive decision support tool in promoting best performance benchmarking and performance modeling. DEA and ANN are combined to take advantages of optimization and prediction capabilities inherent in each method. DEA is used as a preprocessor to measure relative efficiency of decision making units (DMUs) and to generate test inputs for subsequent ANN prediction module. The combined modeling approach proves effective through sequential processes by streamlining DEA analysis and ANN prediction. DEA model, in addition to capturing efficiency trend of the Japanese electronics manufacturing industry, further extends its capacity as a preprocessor to the subsequent neural network prediction module. Back propagation neural network (BPNN), in conjunction with DEA, demonstrates promising performance in predicting efficiency scores and best performance outputs for DMUs under evaluation. This research paper proposes an innovative performance measurement and prediction approach in support of superiority driven best performance modeling and stepwise improvement initiatives.